Research Methods
Possible Final Exam Questions
1, (10 points) Find the standard deviation for the following numbers. Lay out the standard table
and show all computations. The data: 2, 4, 8, 10
2. The z-score formula, the z-test formula and the general form of all statistical tests have corresponding parts. Fill in the blanks (indicated by ?) in the table below to show those corresponding parts (2 points each):
|
z-score: |
z-test: |
general form: |
|
X |
? |
? |
|
? |
0 [zero] |
? |
|
Sx |
? |
? |
3. What (in the context of this course) is meant by "noise" and "signal"? (4)
4. Why is minimizing noise desirable? (4)
5. What are some ways we can minimize noise? (4)
6. Sarah's score on a standardized math test was 72. Give a graphic portrayal of Sarah's z-score
(using a curve with horizontal axis marked off in both z-scores and raw scores, like we did in
class). (The standardized test has a mean of 75, an AD of 4, and Sx of 3, and is normally
distributed). (2)
7. Roughly place the mean, median, and mode in the following skewed curve. (2)
[a curve will be drawn here]
8. Which of these curves, A or B, has a LARGER Sx? Answer: ______
Which has a larger mean? Answer: ___________
[labeled curves will be drawn here]
9. On the axis below, draw
two normal curves that differ only in their means; differ only in standard
deviations; differ in both.
10. Why would it be advantageous to a research project to double the number of people in each
group? Mention EV's, picture of chance, and signal v. noise in your answer. (5)
11. (a) What do we mean by "picture of chance"? (b) How do we obtain a picture of chance?
Describe both the "long way" and the "short way".
12. (12 points) Inferential statistics makes use of information and concepts from throughout this
course. Tell how the following topics covered earlier in the semester are relevant to
inferential stat. Most require only 1-2 sentences. (Note that the request is not for a definition
or explanation. Just tell how the item is RELEVANT to inferential stat):
null and alternative hypotheses
randomization and how it works
RSDDM
TV watching exercise
pictures of chance
active control techniques like holding constant
Sx as a measure of variability
any descriptive statistic (e.g., correlation, mean, difference between two means) is treated as a blur or haze of possible values, rather than as a point
z-scores
normal curve
percentages under the normal curve
signal v. noise
CLT (central limit theorem)
normal curve completely specified by mean and Sx
expected value
13.
Explain why we "fail to reject H0 rather than "accept" H0 .
-----------------------------------------------------------------------
14. Imagine an experiment involving two treatments for depression. Subjects were randomly
assigned to two groups, each group got one treatment for six months. It was hypothesized that
treatment A will relieve depression better than treatment B. At the end of the treatment period, all
were then given the Beck Depression Inventory (measures the degree of depression) to assess
the effectiveness of the treatment.
What is the IV(s)?
What is the DV(s):
What type of design is this?
Name a variable explicitly held constant:
Describe how to match on age:
Describe how to balance gender:
The research hypothesis is:
The null hypothesis is:
This experiment lacks a "control group" that gets no treatment for depression. It also lacks a
"normal control" group of nondepressed subjects. Are these omissions flaws that prevent
testing the hypothesis or lead to confounds? Explain. (6) [Caution: Examine the hypothesis
stated above carefully and ask yourself what groups are needed in order to test it.]
No depression pretest was given to the subjects of this experiment. Does this allow a
confound to arise, or is this a flaw that prevents testing the hypothesis? Explain. (4)
---------------------------------------------------------------------------------------
15. In his famous "bogus pipeline" shock experiment, Milgram examined the effectiveness of
several factors in changing the level of shock subjects were willing to give the "learner". For
example, Milgram either wore an official looking white lab coat over a suit and tie, or just the
suit and tie. Another factor studied was the physical distance between the learner and the
teacher (subject). In one condition they were at adjacent tables, and in another the learner was
in a separate room but could be seen through a one-way mirror. Here are the results, in terms
of average maximum shock "given" by the teachers to the learners (recall that no one was
actually shocked).
no lab coat/separate room: 450
no lab coat/adjacent: 200
lab coat/separate room: 200
lab coat/adjacent: 200
What is the IV(s):
What is the name of this design? ______________________________
What is the DV?
Name a factor explicitly mentioned above that is held constant: ______________________
Interpret the results, using the standard procedure for this design. Draw and label a summary
table and a graph, and make as many statements as necessary. (10)
------------------------------------------------------------------------------
16. In an experiment concerned with the misattribution of arousal in females, two random groups
of 40 women each are given a set of jokes to rate for humor on a 1-7 scale. One group is
"aroused" beforehand by riding an exercize bicycle.
What is the hypothesis of the study (implicit in the design)?
What is the conceptual IV(s)?
What is the operational definition of the IV(s)?
What is conceptual DV(s)?
What is the operational definition of the DV(s)?
What is the design employed -- simple experiment, factorial, quasiexperiment, or correlational?
Is it a flaw in the research to use only women? Explain.
17. Last semester one group of students in this class proposed testing the effect of the V-chip on aggressiveness in children. The V-chip electronically prevents children from watching violent TV programs. We might expect that the V-chip would produce lower aggression than watching violent TV programs, but more aggression (due to the frustration of having a desired program blocked) than not watching TV at all.
18. Imagine the following experiment. A group of boys experience all of the following 3
conditions on different days: watching violent TV, watching TV with violent programs
blocked using the chip, and engaged in playing a nonaggressive/noncompetitive game.
Immediately following the treatment aggressiveness is assessed by counting the number of
physical contacts (hits, kicks, etc.) plus the number of verbal aggressions (yelling,
name-calling, etc.).
What is the hypothesis of the study?
What is the IV(s)?
What is conceptual DV(s)?
What is the operational definition of the DV(s)?
What is the design employed?
Is it a flaw in the research to use only boys? Explain.
What unmentioned control technique MUST be used for the results to be valid?
19. One experiment examined the effects of Prozac in treating internet addiction over a period of
24 weeks. Subjects were 50 college students who spent a minimum of 30 hr./wk on the
internet. Neither they nor the experimenter knew whether each received Prozac or a
look-alike inert pill. Subjects' internet time was monitored using a computer program which
logs internet use.
What is the likely hypothesis of the study?
What is the IV?
What is the inert pill called?
How should the subjects be assigned to the two groups?
What is the name for the fact that neither subject nor experimenter know which treatment is
being received?
Is it a flaw in the design to use only addicts? Why or why not? (4)
What is the DV?
How is the DV operationally defined?
20. Children of alcoholics often show increased susceptibility to stress and fewer coping strategies for dealing with stress than do children of nonalcoholics. Imagine we address this problem by performing the following research:
A group of adult children of alcoholics is divided into two subgroups. The "therapy" group is given a form of group therapy specific designed to teach stress coping skills. The other group does not get this program. At the end of the 3-month program both groups are given the Yates Stress Coping Skills Test (YSCST?), designed to measure the effectiveness of one's coping skills. The hypothesis is that the therapy group will have higher scores, on average, than the other group. (each question except the last is worth 2 points)
How would we go about determining whether the Yates test is reliable?
How could we determine whether the Yates is valid?
If the "other" group receives no treatment at all, this would represent a CONFOUND. Explain.
What should be done to the "other" group to prevent this confounding? (Use the concept of placebo).
Should we let the "other" group know they are a control condition? Why not?
Consider the factor of gender. How would one:
hold it constant?
balance it?
Consider the factor of IQ. How would one match on it?
Could this be done as a "within subjects" experiment? Explain.
Should one randomly assign the subjects to the two groups? Why is this important?
What "level of measurement" is the Yates likely to be? Justify your answer.
Can this experiment be done in a manner that does not violate any of the ethical principles for human subject? Please describe what needs to be done or considered.
(12 points) Assume that the therapy group scored an average of 30 on the Yates, and the other group scored 20. Do a test of significance on this result. Show all work, numbering your steps 1-6. (Note: Sx - y = 3).
21. (a) How do we determine whether a result is "significant"?
This is not a question about computations or steps or mechanics, but a question about the
ambiguous situation we find ourselves in at the end of every experiment, and how we
rationally resolve it by making a basic comparison.
Begin by stating the basic ambiguity of interpretation for any result, then explain what
comparison we make to rationally resolve it. (That is, describe the basic comparison to
chance we make in order to determine whether a result is significant. How is a picture of
chance relevant?)
(b) What two factors do we consider that are taken into account in the picture of chance and
the z-score formula we apply to locate our result in it?
(c) What general form do these considerations give to EVERY statistical test?
(d) What is the role of the number of subjects?
(e) Explain the need for arbitrary cutoff points.
The audience for writing this question is a class member who missed class over the last
couple of weeks. You can assume they already know material from earlier in the semester, but
keep your explanation basic and step by step, and as clear as possible. Audience
appropriateness will be an important consideration in grading your essay. (20 points total)
Various questions can be asked about the following research scenarios, much like the examples above.
22. Are young adults from divorced families more selective/critical in mate selection than those
from nondivorced families? Participants in these two groups responded to an open-ended
question to list 20 characteristics they would require in a potential marriage partner. All
subjects were between 20 - 25, and all were in college or college graduates.
23. A researcher investigated the notion that drinking alcohol affects judgments of attractiveness
so that members of the opposite sex begin to appear more attractive when alcohol is
consumed. Also, males were expected to show more of this effect than females. participants
were tested separately in both alcohol and non-alcohol conditions and their ratings were
compared
24. A researcher hypothesized that sexism and racism occur in the same individuals. Fifty
Caucasian males were given a questionnaire measuring both "isms". it was found that
subjects who scored high on racism also scored high for sexism. Likewise, those who scored
low on racism scored low on sexism. What is the design of this study? Draw a diagram to
indicate the outcome described. Is there an IV? DV? What are the variables being studied.
25. Researchers at the University of Arizona are studying the self-esteem levels among children. They are trying to discover how teachers can best raise the self-esteem levels of their kindergartners. The researchers did a study involving 50 children. The researchers divided the children into two groups. One of the groups received verbal praise for all the activities they did throughout the day, and the other group did not receive this verbal praise. The researchers entitled this verbal praise as self - enhancement. Also, the one group received help with activities they were asked to do, and the other group did not receive help with activities. The researchers entitled receiving help as skill development. At the end of the four-week session, the researchers measured the self - esteem levels of all the children. The researchers hypothesized that the children who received both self-enhancement and skill development would have a higher level of self - esteem than children who received just one or neither of the two help measures. [all of the following questions apply to this scenario]
1.What type of design is this experiment? (Be specific)
2.What is the IV? (If more than one, list all)
3. What is the DV?
4. Name one extraneous variable that should be controlled?
5. Specify how the researchers could employ balancing; randomization; matching (in each case specify an example EV and what operations would be used to control it; e.g., "gender could be held constant by using only females").
balancing:
randomization:
matching:
6. Name one ethical consideration that these researchers need to consider while performing this experiment, and describe how they can go about making sure it is not violated.
7. Illustrate the design of this experiment using a summary table. Fill in the table with your own data that shows the researchers' hypothesis is correct. Graph the results on the back.